A Boolean model with a non-random convex typical grain M is considered
. Set-valued inside and outside estimators for the central symmetrizat
ion M = {x - y: x, y is-an-element-of M} of M are proposed and their c
onsistency in the Hausdorff metric is demonstrated. The estimation met
hod is based on the evaluation of the empirical capacity functional an
d an application of the Gilvenko-Cantelli theorem for random closed se
ts. Generalizations for other models are given (Boolean models with a
non-convex typical grain, multiphased Boolean models, non-stationary B
oolean models etc.).